Nested specific type de-serialization with Avro
apache-kafka, avro, deserialization, java, serialization
Solution
The right way to do it was to use :
GenericData.Array<GenericRecord> productTuple =(GenericData.Array<GenericRecord>) record.get("products");
I believe the last one might work as well but this one lets me deserialize the JSON without any further interference from a class.
Problem
Here is the related part of the schema that I am using to generate my Avro class ``` {"namespace": "com.namespace.kafka.event", "type": "record", "name": "RecordName", "fields": [ // some fields declared ... {"name": "products", "type": { "type": "array", "items": { "name": "productInfo", "type": "record", "fields" : [ {"name": "productId", "type": "int"}, {"name": "productScore", "type": "float"} ] } } }, // some more fields declared ... ``` ] } I am able to generate classes for this schema and to populate them. So basically in the generation a new class called productInfo is generated which serves my needs. (Later I send the data to Kafka). The problem is the de-serialization. I am using something like this: ``` GenericData.Array<String> productTuple = (GenericData.Array<String>) record.get("products"); ``` This returns me a string (which is actually a JSON) like this for each element of GenericData array. ``` {"productName" : someName, "productScore" : someScore } ``` I can parse this string and get what I want but is there a more "Avro" way of doing this. Like can I cast the results again to a GenericRecord object like this. ``` GenericData.Array<GenericRecord> productTuple = (GenericData.Array<GenericRecord>) record.get("products"); ``` or will this work where productInfo being the class generated from my avro schema: ``` GenericData.Array<productInfo> productTuple = (GenericData.Array<productInfo>) ``` Unfortunately, I cant test it I dont have an easy setup to do so. I searched the whole web but could not find a way to do that. Any help will be much appreciated.